The invention relates to apparatus for use in chemical vapor infiltration and deposition (CVI/CVD) processes. More particularly, the invention relates to a preheater for heating a reactant gas inside a CVI/CVD furnace, and a fixture for depositing a matrix within a stack of porous structures by a pressure gradient CVI/CVD process.
Chemical vapor infiltration and deposition (CVI/CVD) is a well known process for depositing a binding matrix within a porous structure. The term "chemical vapor deposition" (CVD) generally implies deposition of a surface coating, but the term is also used to refer to infiltration and deposition of a matrix within a porous structure. As used herein, the term CVI/CVD is intended to refer to infiltration and deposition of a matrix within a porous structure. The technique is particularly suitable for fabricating high temperature structural composites by depositing a carbonaceous or ceramic matrix within a carbonaceous or ceramic porous structure resulting in very useful structures such as carbon/carbon aircraft brake disks, and ceramic combustor or turbine components. The generally known CVI/CVD processes may be classified into four general categories: isothermal, thermal gradient, pressure gradient, and pulsed flow. See W. V. Kotlensky, Deposition of Pyrolytic Carbon in Porous Solids, 8 Chemistry and Physics of Carbon, 173, 190-203 (1973); W. J. Lackey, Review, Status, and Future of the Chemical Vapor Infiltration Process for Fabrication of Fiber-Reinforced Ceramic Composites, Ceram. Eng. Sci. Proc. 10[7-8] 577, 577-81 (1989) (W. J. Lackey refers to the pressure gradient process as "isothermal forced flow"). In an isothermal CVI/CVD process, a reactant gas passes around a heated porous structure at absolute pressures as low as a few millitorr. The gas diffuses into the porous structure driven by concentration gradients and cracks to deposit a binding matrix. This process is also known as "conventional" CVI/CVD. The porous structure is heated to a more or less uniform temperature, hence the term "isothermal," but this is actually a misnomer. Some variations in temperature within the porous structure are inevitable due to uneven heating (essentially unavoidable in most furnaces), cooling of some portions due to reactant gas flow, and heating or cooling of other portions due to heat of reaction effects. In essence, "isothermal" means that there is no attempt to induce a thermal gradient that preferentially affects deposition of a binding matrix. This process is well suited for simultaneously densifying large quantities of porous articles and is particularly suited for making carbon/carbon brake disks. With appropriate processing conditions, a matrix with desirable physical properties can be deposited. However, conventional CVI/CVD may require weeks of continual processing in order to achieve a useful density, and the surface tends to densify first resulting in "seal-coating" that prevents further infiltration of reactant gas into inner regions of the porous structure. Thus, this technique generally requires several surface machining operations that interrupt the densification process.
In a thermal gradient CVI/CVD process, a porous structure is heated in a manner that generates steep thermal gradients that induce deposition in a desired portion of the porous structure. The thermal gradients may be induced by heating only one surface of a porous structure, for example by placing a porous structure surface against a susceptor wall, and may be enhanced by cooling an opposing surface, for example by placing the opposing surface of the porous structure against a liquid cooled wall. Deposition of the binding matrix progresses from the hot surface to the cold surface. The fixturing for a thermal gradient process tends to be complex, expensive, and difficult to implement for densifying relatively large quantities of porous structures.
In a pressure gradient CVI/CVD process, the reactant gas is forced to flow through the porous structure by inducing a pressure gradient from one surface of the porous structure to an opposing surface of the porous structure. Flow rate of the reactant gas is greatly increased relative to the isothermal and thermal gradient processes which results in increased deposition rate of the binding matrix. This process is also known as "forced-flow" CVI/CVD. Prior fixturing for pressure gradient CVI/CVD tends to be complex, expensive, and difficult to implement for densifying large quantities of porous structures. An example of a process that generates a longitudinal pressure gradient along the lengths of a bundle of unidirectional fibers is provided in S. Kamura, N. Takase, S. Kasuya, and E. Yasuda, Fracture Behavior of C Fiber/CVD C Composite, Carbon '80 (German Ceramic Society) (1980). An example of a process that develops a pure radial pressure gradient for densifying an annular porous wall is provided in U.S. Pat. Nos. 4,212,906 and 4,134,360. The annular porous wall disclosed by these patents may be formed from a multitude of stacked annular disks (for making brake disks) or as a unitary tubular structure. For thick-walled structural composites, a pure radial pressure gradient process generates a very large density gradient from the inside cylindrical surface to the outside cylindrical surface of the annular porous wall. Also, the surface subjected to the high pressure tends to densify very rapidly causing that surface to seal and prevent infiltration of the reactant gas to low density regions. This behavior seriously limits the utility of the pure radial pressure gradient process.
Finally, pulsed flow CVI/CVD involves rapidly and cyclically filling and evacuating a chamber containing the heated porous structure with the reactant gas. The cyclical action forces the reactant gas to infiltrate the porous structure and also forces removal of the cracked reactant gas by-products from the porous structure. The equipment to implement such a process is complex, expensive, and difficult to maintain. This process is very difficult to implement for densifying large numbers of porous structures.
Many workers in the art have combined the thermal gradient and pressure gradient processes resulting in a "thermal gradient-forced flow" process. Combining the processes appears to overcome the shortcomings of each of the individual processes and results in very rapid densification of porous structures. However, combining the processes also results in twice the complexity since fixturing and equipment must be provided to induce both thermal and pressure gradients with some degree of control. A process for densifying small disks and tubes according to a thermal gradient-forced flow process is disclosed by U.S. Pat. No. 4,580,524; and by A. J. Caputo and W. J. Lackey, Fabrication Composites by Chemical Vapor Infiltration, Prepared by the OAK RIDGE NATIONAL LABORATORY for the U.S. DEPARTMENT OF ENERGY under Contract No. DE-AD05-840R21400 (1984). According to this process, a fibrous preform is disposed within a water cooled jacket. The top of the preform is heated and a gas is forced to flow through the preform to the heated portion where it cracks and deposits a matrix. A process for depositing a matrix within a tubular porous structure is disclosed by U.S. Pat. No. 4,895,108. According to this process, the outer cylindrical surface of the tubular porous structure is heated and the inner cylindrical surface is cooled by a water jacket. The reactant gas is introduced to the inner cylindrical surface. Similar forced flow-thermal gradient processes for forming various articles are disclosed by T. Hunh, C. V. Burkland, and B. Bustamante, Densification of a Thick Disk Preform with Silicon Carbide Matrix by a CVI Process, Ceram. Eng. Sci. Proc 12[9-10] pp. 2005-2014 (1991); T. M. Besmann, R. A. Lowden, D. P. Stinton, and T. L. Starr, A Method for Rapid Chemical Vapor Infiltration of Ceramic Composites, Journal De Physique, Colloque C5, supplement au n.sup.. 5, Tome 50 (1989); T. D. Gulden, J. L. Kaae, and K. P. Norton, Forced-Flow Thermal-Gradient Chemical Vapor Infiltration (CVI) of Ceramic Matrix Composites, Proc.-Electrochemical Society (1990), 90-12 (Proc. Int. Conf. Chem. Vap. Deposition, 11th, 1990) 546-52. Each of these disclosures describes processes for densifying only one porous article at a time, which is impractical for processing large numbers of composite articles such as carbon/carbon brake disks.
In spite of these advances, apparatus for implementing improved CVI/CVD processes on a large scale are generally desired. Such apparatus would preferably be capable of simultaneously densifying large numbers (as many as hundreds) of individual porous structures in a CVI/CVD process.